
*****************************************************************************
*
* Readme file for
*
* THREE IS A CROWD: INFORMATION AND ELECTORAL COORDINATION IN ARGENTINA
*
* By Adrián Lucardi, Agustín Vallejo and Germán Feierherd
*
* British Journal of Political Science, forthcoming
*
* November 7, 2025
*
*****************************************************************************


*****************************************************************************
* NOTE
*****************************************************************************

- All the analyses were carried out using R version 4.5.2 (2025-10-31) -- "[Not] Part in a Rumble"
- The script was run on a Mac platform with Apple M3 Pro
- To replicate the results, you first need to install and call the following packages and dependencies in R:
install.packages("cowplot")
install.packages("dplyr")
install.packages("fixest")
install.packages("ggcorrplot")
install.packages("ggplot2")
install.packages("ggpmisc")
install.packages("pacman")
install.packages("rdlocrand")
install.packages("rdpower")
install.packages("rdrobust")
install.packages("readxl")
install.packages("tidyr")
install.packages("tidytext")
install.packages("tidyverse")
install.packages("wesanderson")
install.packages("writexl")
install.packages("xtable")


*****************************************************************************
* REPLICATION FILES
*****************************************************************************

1. "code Replication Three_Is_a_Crowd BJPS.R": script used to reproduce the tables and figures reported in the paper and the supplementary materials. This also produces the LaTeX code for the tables, and exports them to the "tables" folder.

2. "Elecciones PBA (by muni) 1983-2023 - Wide.csv": Main dataset used in the analysis. The unit of observation is the municipality-year. For each municipality in each election year, the values for the primary and the general election are reported in the same round.

This dataset is used to replicate the following:
- Table 1
- Figure 3

- Table A1
- Table A2
- Table A3
- Table A4
- Table A5
- Table A6
- Table A7
- Table A8 (jointly)
- Table A9
- Table A10 (jointly)
- Table A11
- Figure A1
- Figure A2
- Figure A4
- Figure A5
- Figure A7

3. "Elecciones PBA (by party) 1983-2023 - Long.csv": Electoral results aggregated at the municipality-year-type of election-party variable. That is, for every municipality-year, there is one observation for each party that participated in the primary, and one observation for each party that took part in the general election. Parties that contested both the primary and the general election thus appear twice per municipality-year.

This dataset is used to replicate the following:
- Figure 1
- Figure 2

4. "Elecciones PBA (by party) 1983-2023 - Wide.csv": Electoral results aggregated at the municipality-year-party. That is, for every municipality-year, there is one observation for each party that participated in either the primary or the general election; for parties that participated in both, the values for the primary and the general election appear in the same row.

This dataset is used to replicate the following:
- Table 2
- Figure 4
- Figure 5

- Table A8 (jointly)
- Table A10 (jointly)
- Table A13
- Table A14
- Table A15
- Table A16
- Table A17
- Table A18
- Figure A3
- Figure A8
- Figure A9
- Figure A10
- Figure A11

5. "Elecciones PBA (by party) 2011-2023 - Provisional.csv": Electoral results aggregated at the municipality-year-type of election-level of election-party variable. That is, for every municipality-year between 2011 and 2023, there is one observation for each party participating at each level of election (municipal, provincial or national) in either the primary or the general election. Parties that contested both the primary and the general election thus appear twice per municipality-year. As explained in fn. 20 in the text, these are provisional rather than definitive results. Nonetheless, in practice both provisional and definitive results are very similar.

This dataset is used to replicate the following:
- Table 3
- Table A12
- Table A19

6. "trends.zip": Original Google trends data on queries and most searched terms. These are used to replicate the following:
- Figure 6
- Figure A6
